BACTERIOLOGICAL STUDY OF RETAIL ICE CREAM. 21 
Table 14. — Cultures picked from Endo's plates showing the number of gas formers . 
Endo plate. 
Colonies 
picked. 
Cultures showing gas in lac- 
tose broth. 
Gas+ 
Gas- 
Gas+ 
1 
Number. 
160 
32 
73 
9 
10 
21 
Number. 
91 
7 
69 
1 

16 
Number. 
69 
25 
4 
8 
10 
5 
Per cent. 
56 87 
21.87 
94.52 
11.11 
.00 
76.19 
9 
3 
4 
5 
6 
From one plate, of the 10 colonies picked off, none produced gas 
in lactose broth. Of 160 cultures from suspected colon colonies from 
sample 1 only 56.87 per cent were gas formers. In sample 3 a high 
percentage — that is, 94.52 per cent of cultures — were gas formers. 
It is evident that in many cases it is impossible to count the typical 
colonies on Endo plates and say that they represent gas formers or 
colon bacilli. Typical colon bacilli are supposed to form on Endo's 
agar medium-sized colonies which are distinctly red and have a 
metallic sheen, but doubtless the appearance of colon colonies is 
quite variable. In a study of the colonies of 10 different strains of 
Bacillus coli on Endo's plates it was found that many of the colonies 
showed no metallic sheen. Some were pinkish instead of red; there 
was a red coloration of the medium around some colonies and none 
around others, and many of the colonies were very small. We also 
found that a number of nongas-forming bacteria of the acid group 
and some peptonizers would grow readily on Endo's medium, and 
the colonies had all the appearance of those of some of the 10 strains 
of B. coli used during these experiments. It must be conceded, 
however, that a much larger number of gas-forming bacteria can be 
found in ice cream when using Endo's medium than when plating 
on'litmus-lactose-asparagin agar, and the asparagin plates showed a 
much higher number than the lactose-bile tubes. The use of Endo's 
medium for the determination of colon bacilli in ice cream is com- 
plicated by the fact that bacteria other than gas formers may give 
the typical reaction of the colon bacilli; also because when a dilution 
low enough to determine the number of B. coli is used the plate is 
sometimes overgrown by other bacteria which are present in large 
numbers. 
It is not the purpose of this paper to discuss the best method for 
the determination of B. coli in ice cream, for the reason, which is 
obvious, that an entirely satisfactory method has not yet been per- 
fected. Each process has its merits and objectionable features, but 
in view of our present knowledge of the colon group of bacteria it is 
impossible to say that any one method is the best. 
